update unit conversion message responses in plugins, add tests

[cshanahan1]

This PR improves the messaging from the unit conversion plugin to broadcast changes in selected display unit. Because some plugins are listening for changes to the spectral y axis (either from toggling flux<>sb, changing flux or angle when it is toggled to SB, or changing flux when it is toggled to flux) this is its own message now.

Every time a new selection of 'flux' is made, there are three GlobalDisplayUnitChanged messages broadcasted : 'flux', 'sb', and 'spectral_y'.

While this is not included in this PR, these changes will make it possible for the correct messaging to be broadcasted when a new angle selection is made (always sending a 'sb' message, and if the spectral y axis is in SB also sending a 'spectral_y' message).

Every time there is a toggle between flux<> sb, the GlobalDisplayUnitChanged axis='spectral_y' message is broadcasted.

This PR also updates the plugins listening to unit changes to intercept only the relevant type of GlobalDisplayUnitChanged - before, they were responding to any message of this type.

Finally, this fixes an issue with mouseover in cubeviz and adds test coverage.

[pllim]

Is a regression test easy to add? Both you and Ricky seemed surprised that CI didn't catch this.

[codecov]

Codecov Report

Attention: Patch coverage is 66.66667% with 1 line in your changes missing coverage. Please review.

Project coverage is 88.80%. Comparing base (f5c93ab) to head (42d5bae).

Files	Patch %	Lines
...specviz/plugins/unit_conversion/unit_conversion.py	50.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3144      +/-   ##
==========================================
- Coverage   88.82%   88.80%   -0.02%     
==========================================
  Files         112      112              
  Lines       17550    17552       +2     
==========================================
- Hits        15588    15587       -1     
- Misses       1962     1965       +3     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[kecnry]

this probably needs some logic to skip if the msg.axis is not either spectral or spectral_y (could be by checking if axis is None after the get, or rewriting this into if, elif, else: return.

[kecnry]

Is there any reason this can't use the self.flux_or_sb switch instead of self.spectrum_viewer.state.y_display_unit (maybe @gibsongreen would have some context on that as well)? That could be done as a follow-up if its out of scope.

Either way, maybe we rename the variable to spectral_y to be consistent with the message that is broadcast from this and to avoid confusion with the switch itself?

[gibsongreen]

I think that should work, the traitlet for self.flux_or_sb at this point in the function should still match that flux or surface brightness selection for the value of self.spectrum_viewer.state.y_display_unit/spectral_y/current_y_unit